A simulation-first integration prototype that connects a Python finite-state controller to PX4 SITL and Gazebo Classic through MAVROS. The project explores a shepherding-style multi-agent scenario; it is not a flight-tested autonomous system.
- PX4 SITL
v1.14.0-rc1 - Gazebo Classic 11
- ROS Noetic and MAVROS
- Python 3.8 controller
The controller uses a five-state finite-state machine to coordinate high-level behavior:
- initialize and connect;
- arm and take off;
- approach the simulated target area;
- execute the shepherding behavior;
- return or terminate safely in simulation.
The included figure documents the intended simulation integration. The work demonstrates SITL, ROS, MAVROS, and Python-controller integration. It does not claim real-world flight performance, vision-detector completion, collision safety, or field validation. See docs/evidence.md.
assets/ Public simulation overview
docs/ Evidence scope and reproducibility boundary
This is a portfolio case study, not a ready-to-fly vehicle package. Rebuild the simulation with your own PX4, Gazebo, ROS, and vehicle configuration before extending it.
Apache-2.0. See LICENSE.